Tires are also a 1 to 3% variable on a dyno. Tires can "rob power", inflation psi, traction, tread design. One thing is for sure, when you change something (rear gear ratio, trans ratio, tires, even
oil if a different type AND at a different temperature), you need a new base line before you begin tuning again.
These variables are the big reason it's impossible to say how much percent of parasitic loss you can expect from one car to the next. Could be 17%, could be 25%, with seemingly duplicate spec cars!
A chassis dyno is liking weighing yourself with your clothes on. Wet cloths? Dry cloths? Empty pockets? Boot, shoes or barefoot?